Intensification of Continuous Ortho-Lithiation at Ambient Conditions - Process Understanding and Assessment of Sustainability Benefits

Article abstract of DOI:10.1021/acs.oprd.7b00142

The feasibility of performing an ortho-lithiation reaction in a T-reactor and a spinning disc reactor (SDR) at ambient temperature has been demonstrated and compared experimentally to a conventional batch stirred tank reactor (STR) process performed at a

Full text of DOI:10.1021/acs.oprd.7b00142

Subscriber access provided by UNIVERSITY OF CONNECTICUT
Full Paper
INTENSIFICATION OF CONTINUOUS ORTHO-
LITHIATION AT AMBIENT CONDITIONS – Process
Understanding and Assessment of Sustainability Benefits
Ruili Feng, Sushil Ramchandani, Balamurugan Ramalingam, Song Wei Benjamin
Tan, Chuanzhao Li, Soo Khean Teoh, Kamelia Boodhoo, and Paul N. Sharratt
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.7b00142 • Publication Date (Web): 07 Jul 2017
Downloaded from http://pubs.acs.org on July 7, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Organic Process Research & Development is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
INTENSIFICATION OF CONTINUOUS ORTHO-
LITHIATION AT AMBIENT CONDITIONS –
Process Understanding and Assessment of
Sustainability Benefits
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a,b
b
c
b
Ruili Feng, Sushil Ramchandani, Balamurugan Ramalingam, Song Wei Benjamin Tan,
a
b
a*
b
Chuanzhao Li, Soo Khean Teoh, Kamelia Boodhoo and Paul Sharratt
a
School of Chemical Engineering and Advanced Materials, Newcastle University, Newcastle
Upon Tyne, NE1 7RU, United Kingdom
b
INCOME team, Process Science and Modelling group, Institute of Chemical and Engineering
Sciences (ICES), 1 Pesek Road, Jurong Island, Singapore 627833
c
Organic Chemistry group, Institute of Chemical and Engineering Sciences (ICES), Neuros 8
Biomedical Grove Neuros Building, Singapore 138665
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table of Contents Graphic
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
KEYWORDS Process intensification, lithiation, sustainability, spinning disc reactor, continuous
processing, microreactor
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ABSTRACT The feasibility of performing an ortho-lithiation reaction in a T-reactor and a
Spinning Disc Reactor (SDR) at ambient temperature has been demonstrated and compared
experimentally to a conventional batch stirred tank reactor (STR) process performed at a
cryogenic temperature of -70 ˚C. Benefits of significantly improved mixing and much shorter
residence times in the flow reactors eliminated the need for cryogenic cooling in batch
processing which is a costly requirement. A theoretical evaluation of a scaled up process with a
design product output of 3 tons per year highlights that significant process intensification is
achievable in the flow reactors which demonstrate higher energy efficiency, better volume
efficiency, smaller processing inventory and smaller equipment footprint. For this reaction, the
performance of the T-reactor is the best among the three reactors leading to much lower reactor
investment and operating cost. The SDR was also demonstrated to be effective, even though this
reaction did not take advantage of its inherent capabilities in solids handling or rapid heat
removal through evaporation/gas disengagement, all of which present major challenges in
enclosed channels geometries such as the T-reactor.
1
. Introduction
Ortho-lithiation is an important class of reaction for the synthesis of regiospecifically
substituted aromatics and it is an emerging method to prepare phthalides which are common
1
-6
pharmaceutically active compounds. Ortho-lithiation is typically conducted in batch mode
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
under cryogenic temperatures (-78 to -40 ˚C) to minimize the side reactions arising from the
highly reactive organlithium intermediates. The challenges of safely handling both the highly
reactive chemicals and a highly energetic reaction in scaled-up operations constitute a major
disadvantage of such batch processes. One solution to overcome these safety related issues is to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
8
implement continuous flow processes.
Indeed, several successful studies have been performed to scale up continuous lithiation using
4
, 6
tubular flow reactors. These flow reactors offer superior control of process parameters like
mixing, residence time and temperature. There are examples of deploying flow processing for
lithiation reaction as a process intensification (PI) technique. The Yoshida group demonstrated
several lithiation reactions using different electrophiles in microstructured flow devices
9
-13
(consisting of micromixers and microtube reactors) to be superior to batch protocols.
Roberge
5
4
and co-workers and Stephan and co-workers have both demonstrated the possibility to scale up
lithiation reactions in flow reactors for pharmaceutical manufacturing. However, both authors
reported challenges in handling plugging due to precipitation of salts. Kockmann and co-
14
workers evaluated the feasibility to scale up the lithiation reaction using tubular flow reactors
with different channel diameters and flow rates to avoid parallelization.
One major problem with adopting flow processing in enclosed microchannels is the possibility
4
of deposition of the intermediate lithium salts, which can lead to clogging of such flow reactors.
We envision overcoming this limitation by the use of the SDR which allows a free surface film
flowing over the disc surface instead of through an enclosed channel. The SDR also possesses a
number of characteristics which make it a potential intensification tool for ortho-lithiation
reaction, such as the capability of achieving rapid mixing with an estimated micromixing time in
15
the range of 0.125 to 0.02 s and high mass and heat transfer rates in the free flowing thin film
4
ACS Paragon Plus Environment

Page 5 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
16
of liquid produced owing to centrifugal acceleration created by rotation. Various examples in
17-19
the literature demonstrate the advantages of performing fast reactions in the SDR.
Inspired
by these advantages, the feasibility of performing the ortho-lithiation reaction was demonstrated.
The objective of this work is to evaluate the sustainability benefits of operating ortho-lithiation
reaction in different process intensification (PI) reactors as compared to the conventional batch
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
0
reactor. In this work we focus on comparing the sustainability performance of a batch and two
intensive continuous reactors (the T-reactor and the spinning disc reactor), for the reaction step
excluding workup at a hypothetical design scale of 3 tons per year. The potential benefits that
could be achieved are higher reaction selectivity and material efficiency, lower energy
consumption, improved safety and economic savings.
2
2
. Experimental work
.1. Reaction system
Scheme 1. Reaction used in this study and possible side reactions.
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
A typical ortho-lithiation reaction is taken from the literature (Scheme 1), where the 4-chloro-
N,N-diisopropylbenzamide (amide 1) is treated with n-Butyllithium (n-BuLi, 1.6 M in hexane) in
tetrahydrofuran (THF) at -70 to -75 ˚C and the lithiation occurs selectively at the ortho-position
to the amide group. The subsequent treatment with dimethylformamide (DMF) in Step 2
produces 4-chloro-N,N-diisopropyl-2-formylbenzamide (product 1). Step 1 is a very fast reaction
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
1
with an adiabatic temperature rise of more than 55 ˚C (230 kJ/mol n-BuLi). The second step is
less demanding in terms of heat exchange and mixing and has a lesser possibility for new side
1
4
product formation.
Unfortunately the side reaction mechanism of ortho-lithiation is not yet fully understood. Only
partial information was obtained in our study using Liquid Chromatography-Mass Spectrometry
(LC-MS) which nevertheless provided some useful insight into the possible side products
formed. As shown in Scheme 1, two possible side reactions have been proposed based on
identification by LC-MS of two molecular masses which have been matched to our anticipated
molecular structures SP 1 and SP 2. It is to be noted that versions of other structural isomers of
these molecules are also possible. It is postulated that these side products arise from unselective
lithiation reactions competing with the desired ortho-lithiation in Step 1. The mixing
performance in Step 2 determines the formation of either SP 1 or SP 2. Poor mixing in Step 2
tends to generate SP 1 as DMF fails to react with the lithiated species before contacting water,
while good mixing in Step 2 allows the lithiated species to undergo electrophilic addition to form
SP 2. Certain processing conditions may also affect the overall amide conversion. For instance,
2
2
n-BuLi tends to decompose or get consumed in the reaction with THF when n-BuLi is not well-
mixed. A lesser quantity of n-BuLi available for reaction with amide 1 leads to low overall
conversion. Similarly low overall conversion is expected when the DMF fails to mix
6
ACS Paragon Plus Environment

Page 7 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
homogeneously with the reaction mixture in Step 2, so the intermediate is converted back to
amide 1 upon contact with water during separation process.
Three types of reactors were evaluated for sustainability comparisons:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
) Batch stirred-tank reactor (STR) under cryogenic conditions – conventional manufacturing
process.
) T-reactors operating in flow at ambient temperature – similar to microreactor but with
larger channel diameters to avoid clogging.
3) Spinning disc reactor (SDR) in flow at ambient temperature – giving a very different mode
of mixing and operation from the T-reactor.
2.2. Experimental procedures
For the stirred tank reactor, a 50 mL Multi-max™ reactor (with standard overhead stirrer, 2
blade propeller impeller with diameter of 2.5 cm) was employed to conduct the ortho-lithiation
reaction over a range of temperatures between -75 and 20 ˚C and impeller agitation speed
between 50 to 400 rpm. A total of 2 g (8 mmol) of amide 1 was added into the reactor followed
by 19 g of solvent, THF (anhydrous, 99.8%). The n-BuLi in hexanes (1.6 M, 1.2 mole
equivalent) was added dropwise (0.4 mmol/min) via syringe pump (NE 1000) to the solution of
starting material in THF in the reactor. The reactor was submerged in an isopropyl alcohol (IPA)
bath with constant addition of dry ice to maintain reactor internal temperature at about -70 ˚C
monitored using a digital temperature sensor in contact with the reaction mixture. After which
the reaction mixture was allowed to warm to about -50 ˚C when DMF (0.8 g, 11 mmol, 1.3 mole
equivalent) was added via syringe pump (NE 300). During the reaction, an inert reaction
environment was maintained under argon blanket.
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The preparation steps for both the T-reactors and the SDR were similar. 0.4 M of amide 1
solution was prepared by dissolving amide 1 (24 g) in THF using a 250 mL volumetric flask.
About 120 mL of the amide 1 solution, 40 mL of n-BuLi in hexanes (1.6M) and 10 mL of DMF
were drawn using syringes (HSW syringes) and loaded onto syringe pumps (NE 1010 dual, NE
1000, NE300 respectively). Special care was taken to ensure the drawing of n-BuLi into the
syringe was not exposed to air. The T-reactor was flushed and filled with anhydrous THF for at
least 10 minutes before experiments to ensure an anhydrous environment. Similarly for SDR, the
reactor was purged with argon prior to the experiment and kept under positive argon pressure
throughout experiment. The collected solution was sent to waste for the first 30 s to allow the
system to reach steady state. After this time the output of the reactor was collected in fixed time
interval of about every 30 s. However, due to limited volume of the syringe, the flow reactors
could only be operated for a short time (about 5 minutes). Samples were taken at different time
intervals to ensure steady state has been reached.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The T-reactor was assembled in house from commercially available PTFE tubing (Sigma
Aldrich, 1.6 mm I.D.) and connected to of two PTFE T-mixers (Upchurch Scientific P-713, 1.25
mm I.D., 17.5 µL swept volume). The T-reactor setup, included a reactor tubing, R1, ranged
from 0.06 to 2 m. The n-BuLi in hexanes and the amide 1 in THF streams were pumped into the
first T-mixer from opposite directions and the DMF stream merged with the reaction mixture in
the second T-mixer. Other than the syringe pumps and the outlet tubing, the T-reactor was
immersed in a circulated silicone oil bath using Huber temperature control unit.
For the SDR, the stainless-steel disc of 10 cm in diameter was driven by an electric motor
which operates in the range of 400-2,400 rpm. The SDR was tightly enclosed and purged with
argon prior to the experiment and kept at positive argon gas pressure throughout the experiment.
8
ACS Paragon Plus Environment

Page 9 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Cooling water was recirculated underneath the disk surface through a temperature-controlled
water bath to keep the disc temperature constant. The three reagents were delivered via syringe
pumps onto the rotating disk through three feed pipes. The two center feed pipes delivering
amide 1 in THF and n-BuLi in hexanes were located at the center of the disc. The third feed pipe
delivering DMF was located at a radial distance of 45 mm from the center of the disc. All the
feed pipes were fixed at a distance of 5.0 mm above the surface. The feed tube diameter was
approximately 1.6 mm for the starting material stream and 1.0 mm for the n-BuLi in hexanes and
DMF streams where the flow rates were slower. The reactor diagram is depicted in Figure 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(a)
(b)
(c)
Figure 1 (a) Schematic of the SDR (10 cm diameter) set-up; (b) view of disc surface; (c) picture
of the set-up.
23
The n-BuLi (~1.6 M solution in hexanes) was titrated once a week. It is noted that the
concentrations of the n-BuLi may vary (1.4-1.6 M). To account for the variation in n-BuLi
concentration, the corresponding n-BuLi flow rates were adjusted.
2
.3. Method of analysis
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The internal standard method was used to analyze samples collected at fixed time intervals.
The product 1 yield and overall conversion were determined by Gas Chromatography (GC) using
Agilent (GC 6890N) GC-FID equipped with HP-5 column. Response factors of the product 1
and amide 1 were determined by calibration using n-hexadecane as the internal standard.
LC-MS analysis was done at ambient temperature via reverse phase liquid chromatography
with Tandem Mass Spectrometry (Waters Quattro Micro API ™ LCMS/MS) with Restek Ultra
C8 column (5 μm, 4.6 mm (ID), 150 mm (L)) an isocractic elution gradient of 50%
acetonitrile/water was used. The samples were ionized by ESI and the ionized components were
analysed by a quadrupole analyzer. Retention time of amide 1 is 27 min; product 1 is 19 min; SP
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3
3
is 15 min; SP 2 is 10 min.
. Results and discussions
.1. Development of stirred-tank reactor (STR) batch process
3.1.1. Effect of temperature
Temperature is one of the most important process parameters in for ortho-lithiation reaction in
batch process. Attempts were made to perform the reaction at higher temperatures but side
products were formed in large extend as shown in Table 1, the drastic drop in yield. As expected,
-70 ˚C is necessary for reaction in stirred tank reactor. In large scale batch production, the
addition rate of n-BuLi should be adjusted to maintain the reaction temperature below -70 ˚C.
Table 1 Ortho-lithiation at different temperatures at constant agitation speed of 400 rpm in batch
process.
Temperature (˚C) Product
1 Overall
conversion Impurity* (%)
90
yield (%)
(%)
93
2
0
3
1
0
ACS Paragon Plus Environment

Page 11 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
0
26
36
96
99
97
99
73
61
3
-
-
(
20
70
*) % impurity = % overall conversion - % product 1 yield
3
.1.2. Effect of agitation speed
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Mixing is also very important for ortho-lithiation which is a fast reaction. As seen in Table 2,
the results are very dependent on mixing. Poor mixing not only lead to increase in impurities but
also lower overall conversion. The low agitation speed was expected to lead to a high n-BuLi
localized concentration due to ineffective mixing which might encourage decomposition and/or
2
2
attack on THF. The loss of n-BuLi is evident from the low consumption of amide 1 as seen in
Table 2 at agitation speeds of 50 and 100 rpm.
For the Multi-max™ reactor used, the maximum agitation speed achievable was 400 rpm. This
corresponds to a power dissipation of 0.053 W/kg which is set as the minimum power dissipation
for batch large scale production
Table 2 Ortho-lithiation at 0 ˚C under different agitation speeds where n-BuLi and DMF addition
rates were kept constant at 0.4 and 2 mmol/min respectively.
Agitation (rpm)
Product
yield (%)
1 Overall
conversion (%)
Impurity* (%) Power (W/kg)
400
200
26
25
22
7
99
99
98
42
36
73
74
76
35
29
0.053
0.007
0.003
0.0008
0.0001
1
1
5
50
00
0
7
(
*) % impurity = % overall conversion - % product 1 yield
3
.1.3. Effect of DMF addition mode in Step 2
1
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
To study the effect of mixing on Step 2, two different DMF addition rates were used while
keeping the agitation speed of the reactor constant at 400 rpm (Table 3). The high charging rate
of DMF might have caused inhomogeneous mixing and hotspot formation that could lead to
decomposition of the intermediates, where some of the lithiated intermediates might have been
converted back to amide 1. This explains the lower overall conversion of amide 1 at fast DMF
dosing rate. The product in the crude reaction mixture (after DMF addition) is relatively stable
for up to 1 hour prior to the work-up process.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 3 Addition mode of DMF at -50 ˚C (Step 1 performed at -70 ˚C)
DMF
rate
addition Product 1 yield Overall conversion Impurity*
Agitation
speed (rpm)
400
(%)
(%)
63
(%)
33
One shot (~10 30
mmol/s)
Usual dosing rate 96
99
3
400
(
2 mmol/min)
(*) % impurity = % overall conversion - % product 1 yield
3
.2. Development of T-reactor flow process
3.2.1. Effect of DMF addition mode in Step 2
The addition mode/mixing of DMF in Step 2 plays an important part in determining the overall
product 1 yield and overall amide 1 conversion. It is evident from Table 4 that the DMF flow
rate can influence the product 1 yield. Beyond a ‘minimum’ DMF quantity, the mixing limitation
in Step 2 diminishes and reaction profile depends solely on the flow conditions in Step 1.
However, the ‘minimum’ DMF quantity is likely to be a variable which is dependent on Step 1
flow rate.
For simplicity of scale up and material efficiency, DMF of 2 mole equivalent with respect to
amide 1 was used in all flow experiments. For T-reactor, the same mixing principles in Step 1
1
2
ACS Paragon Plus Environment

Page 13 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
can be applied to Step 2 as the reaction mixture meets ‘head on’ with the DMF feed in the
second T-mixer. The speeds of the feeds in the second T-mixer mixing channel are dependent on
the combined flow speed in the first T-mixer. Thus, it is expected that mixing time in the second
T-mixer is of similar order of magnitude as the first T-mixer. Therefore, good mixing
performance in the first T-mixer is likely to improve mixing in the second T-mixer. It is also
noted that poor reaction performance could be a result of poor mixing contributed by both steps.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4 Experiments using different quantities of DMF carried out under constant Step 1
residence time of 0.4 s and bath temperature set at 17 ˚C.
Run
DMF flow DMF
Step
1
Step
combined
2
Product
yield (%)
1
Overall
conversion
(%)
Impurity*
(%)
rate (mL/s)
mole ratio combined
to amide 1 flow
rate flow
(mL/s)
rate
(
mL/s)
1
2
3
4
5
6
0.01
0.02
0.03
0.05
0.08
1.3
1.7
2
0.65
0.65
0.65
0.65
0.65
0.65
0.66
0.67
0.68
0.70
0.73
NA
74
90
92
91
90
91
97
99
99
99
99
99
23
9
7
4
8
6
9
DMF
~
8
reservoir
unlimited
(*) % impurity = % overall conversion - % product 1 yield
3
.2.2. Effect of circulation bath temperature
Attempts were made to change the reaction temperature of the T-reactor by adjusting the bath
temperature. Figure 2 shows that the reaction seemed robust over a range of bath temperature
from 0 ˚C to 40 ˚C. Assuming that the system is adiabatic as PTFE is a good insulator, the
temperature rise is expected to be more than 50 ˚C and yet there was no significant deterioration
2
4-25
in product 1 formation observed in Figure 2. This is referred to as Flash Chemistry
where
reaction is usually completed in less than a second coupled with highly efficient mixing. For this
1
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
type of condition, the mixing performance is more important than temperature control as the
residence times are very short so side reactions due to temperature rise become less importance.
If this is true, then the poor heat transfer capability of the T-reactor may be compensated by
improving the mixing performance and shortening the residence time. The thermal stabilities of
the product 1 and amide 1 up to 300 ˚C were determined using differential scanning calorimetry
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(DSC) and no secondary decomposition is observed. However, it was difficult to conduct
thermal stability test on the intermediate as they are unstable and decompose instantaneously in
air. The circulated silicon bath temperature was maintained at 17 ˚C for subsequent experiments.
1
00
9
9
8
8
7
7
6
6
5
0
5
0
5
0
5
0
Product 1 yield
Overall conversion
0
10
20
30
40
50
Temperature (C)
Figure 2 Effect of circulation bath temperature on Step 1 under constant residence time (0.4 s)
and combined flow speed.
3
.2.3. Effect of residence time in Step 1
Figure 3 shows the effect of residence time for two different flow speeds and it is
demonstrated that shorter Step 1 residence time improves the product 1 yield. The longer than
desired residence time could result in poorer yield due to prolonged exposure to high temperature
which might cause decomposition of intermediate or overreaction.
One observation made from Figure 3 (a) is that the gap in product 1 yield between Reynolds
numbers (Re) of 1219 and 2437 appears more significant at shorter residence times. One possible
reason could be the effect of mixing limitation is more prominent at low flow speed (Re=1219)
1
4
ACS Paragon Plus Environment

Page 15 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
and short residence times (below 0.8 s). As the residence time increases for Re of 1219, there is a
subtle improvement in product 1 yield as there is more time available for mixing to occur until
over-reaction kicks in as characterized by the drop in yield. Therefore, the optimum residence
time is very much mixing dependent.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
100
1
00
95
90
85
80
75
70
65
60
55
50
95
90
85
80
75
70
65
60
55
50
Re=1219
Re=2437
Re=1219
Re=2437
0.100
1.000
10.000
0.100
1.000
10.000
Residence Time (s)
Tres (s)
(
3
a)
(b)
(a) Product 1 yield and; (b) overall conversion against residence time (Step 1) in first T-mixer
for different Re numbers of 1219 and 2437.
3.2.4. Effect of flow speed in Step 1
Figure 4 shows the effect of flow speed in Step 1 on reaction profile for two different residence
times, at 0.4 s and 0.8 s. The range of Reynolds number achieved in the T-mixer mixing channel
varies from 300 to 9800. The geometry of the T-mixer stays constant for various Re but the
subsequent PTFE tube that connects to the T-mixer outlet varies in length to maintain constant
overall residence time.
In the regime where Reynolds number is lower than 2400, the reaction is mixing controlled as
the flow in the pipeline is likely to be laminar where the velocity vectors are parallel and there is
little radial mixing. These laminar flows introduce gradients in reagent concentrations and
temperature as the reaction between n-BuLi and amide 1 occurs in the interface of the two
parallel streams due to molecular interaction arising from diffusion. The bulk of amide 1
1
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
remains unreacted as the two reagents are not homogeneously mixed while the rest of the n-BuLi
gets consumed by reacting unselectively with THF and DMF that are present in abundance. This
explains the low overall conversion at the lowest flow rate. As flow speed increases, it is likely
that the flow in the T-mixer transitions towards turbulence and the resulting enhanced mixing of
the two fluid streams generates increased contact area, higher product 1 yield (Figure 4 (a)) and
overall conversion (Figure 4 (b)).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Under the same flow speed as in the mixing limited regime, longer residence time provides
more time for mixing/diffusion which could result in higher overall conversion. This is evident at
Re 609 where the overall conversion at 0.8 s is clearly higher than that at 0.4 s. As Re increases,
the corresponding mixing limitation diminishes and the benefit of having longer residence time
becomes negligible as the overall conversion approaches 100%. An optimum yield/selectivity is
obtained at Re 2437 as the product 1 yield/selectivity at 0.4 s surpasses that at 0.8 s and this
could be the tipping point where the controlling factor transits from mixing to residence time.
As the Re increased beyond 2400, the product 1 yield/selectivity reached a plateau or is
slightly decreased. One possible reason could be the back-mixing effect in the n-BuLi feed arm
as the amide 1 feed speed is more than three times higher than the n-BuLi. The pressure
difference between the feed inlets gets more severe as flow speed increases. Back-mixing could
cause more side reactions to occur in the n-BuLi feed arm where localized n-BuLi concentration
is high, which accounts for the lower yield/selectivity observed at Re above 4800. Similar
26
observation is made by Kockmann and co-workers where the segregation index is increased for
similar range of Reynolds number used in T-mixer.
1
6
ACS Paragon Plus Environment

Page 17 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
1
00
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
9
8
7
6
0
0
0
0
50
40
30
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
0
0
0
0
.4 s Step 1 residence time
.8 s Step 1 residence time
0.4 s Step 1 residence time
.8 s Step 1 residence time
10.0
.0
1
0
0
0
100
1000
10000
1
00
1000
10000
Reynolds number in T-mixer
Reynolds number in T-mixer
(b)
(
a)
1
00.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0
.4 s Step 1 residence time
0.8 s Step 1 residence time
0
.0
1
00
1000
10000
Reynolds number in T-mixer
(c)
Figure 4 (a) Product 1 yield; (b) Overall conversion; (c) Product 1 selectivity against Step 1
combined flow speed in terms of Reynolds number in the first T-mixer.
3
.3. Development of the SDR flow process
3
.3.1. Mode of DMF addition
The DMF feed is added close to the edge of the disc (4.5 cm radial distance away from the
center of the disc) where the mixing performance should be better compared to other parts of the
disc. While the residence time of Step 2 on the disc is expected to be even shorter than Step 1 as
it has only a radial distance of 5 mm to mix and react, Step 2 reaction is expected to be faster
1
4
than Step 1. Thereafter, the reaction mixture flies off the disc and splashes onto the housing
1
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
wall which is, in fact, a form of mixing as well. Theoretically, SDR offers relatively better
mixing for Step 2 reaction than Step 1 provided that the DMF feed contacts the disc as a
continuous stream rather than dropwise. Through visual observation, a minimum flow rate of 2
mole equivalent (0.06 mL/s) of DMF is required for the lowest Step 1 combined flow rate of 1.3
mL/s to ensure a continuous stream of DMF from the feed pipe.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
.3.2. Position of n-BuLi feed
Theoretically, the mixing of the amide 1 and n-BuLi feeds could be improved by moving one
of the feed (e.g. n-BuLi) away from the center of the disc, where the mixing is poor, to a higher
15
radial position where the film thickness is smaller. However, the result obtained for n-BuLi
feed placed 2.5 cm from center of the disc were worse than position at center of disc (Table 5).
The poorer overall conversion could be explained by the reduction in residence time as the n-
BuLi feed moves 2.5 cm closer to the edge of the disc. The increase in impurity might be the
result of increased n-BuLi to amide 1 stoichiometric ratio as the n-BuLi feed stream is not evenly
distributed over the entire disc but expected to be concentrated at the point of contact. As
2
7
demonstrated by Boiarkina in dye experiments with well-defined and minimal spreading path
lines of the injected dye across the radius of the disc. It is evident from Table 6 that higher n-
BuLi stoichiometric ratio increases formation of impurity.
Table 5 Effect of positon of n-BuLi feed on reaction profile
Position of Flow
n-BuLi feed rate
Disc
speed
(rpm)
Product 1 Overall
Impurity* (%)
10
yield (%)
conversion
(mL/s)
(%)
Center
disc
of 1.3
1400
87
97
1
8
ACS Paragon Plus Environment

Page 19 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
2
.5 cm away 1.3
1400
66
90
24
from center
of the disc
(*) % impurity = % overall conversion - % product 1 yield
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 6 Flow rate of 1.3 mL/s and disc speed of 1400 rpm
n-BuLi mole eqv
Product 1 yield (%)
Overall conversion Impurity* (%)
%)
(
0
.9
1.3
.6
84
87
77
85
99
99
1
12
22
1
(
*) % impurity = % overall conversion - % product 1 yield
.3.3. Effect of SDR coolant temperature
The coolant temperature dictates the disc surface temperature in the SDR. Figure 5 shows that
3
the yield is only marginally affected by the coolant temperature. Similar explanation as in the T-
reactor can be applied where the effect of temperature is limited by the short contact time
between the reagents and the disc. Although compared to the T-reactor, the SDR is expected to
have a better heat transfer capability where the coolant is supplied directly beneath the metal
rotating disc which has a relatively large surface area for heat transfer, the SDR has even shorter
residence time and thus shorter exposure time to the higher surface temperature than the T-
reactor.
1
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
00
9
0
80
7
6
5
4
0
0
0
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
30
Product 1 yield
2
0
0
0
Overall conversion
Product 1 selectivity
1
0
10
20
30
40
50
Temperature (C)
Figure 5 Effect of circulated bath temperature under constant residence time (0.4 s), total flow
rate (1.2 mL/s) and disc speed (1,400 rpm).
3
.3.4. Effect of total flow rate (Step 1)
Figure 6 shows a steady decrease in the product 1 yield and overall conversion with the
increase in flow rate from 1.2 to 5 mL/s at a disc speed of 1,400 rpm.
1
00
8
0
60
40
20
0
Product 1 yield
Overall conversion
Product 1 selectivity
0
1
2
3
4
5
6
Flow rate (mL/s)
Figure 6 Effect of total flow rate (Step 1) under constant disc speed of 1,400 rpm.
Under the same disc speed, Figure 7 shows the corresponding residence time for the flow rates
in Figure 6 at 1,400 rpm. The residence time decreases with increasing flow rates, from 0.15 s at
lowest flow rate to 0.05 s at the highest flow rate. It is known that ortho-lithiation reaction time is
2
0
ACS Paragon Plus Environment

Page 21 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
very fast especially at high temperature, so the reaction is most likely mixing limited in this
1
5
regime. The micromixing time reported by Hinrichsen using similar operating conditions and
disc size, is in the range of 0.125 to 0.02 s. If a conservative estimate of 0.125 s is taken to be the
average micromixing time regardless of flow rate and disc speed, Figure 7 shows that for disc
speed of 1,400 rpm and flow rates above 1.2 mL/s, the corresponding residence times could be
shorter than the micromixing time (represented by the dotted line). This suggests that the mixing
controlled reaction might be indirectly limited by the short residence time. Therefore, the longer
residence time benefits the mixing as it allows more time for molecular contact. This trend is
similarly observed in the T-reactor with low flow rates.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0
0
0
0
.40
.35
.30
.25
1ml/s
2 ml/s
3ml/s
4 ml/s
5ml/s
0.20
0.15
0.10
0
.05
.00
0
0
500
1000
1500
2000
2500
3000
Rotational speed (rpm)
2
8
Figure 7 Estimated residence time of Step 1 on the disc at different flow rates and disc speeds.
Data within the pink area represents the corresponding residence times at 1,400 rpm.
Although the SDR residence time can be extended by reducing the disc speed, the mixing
intensity might be compromised even further. Compared to the T-reactor which offers residence
times of 0.2 to 3 s, the estimated range of residence time of 0.05 to 0.15 s used in the existing
SDR is much shorter. The shorter residence time on the SDR is more likely to minimize over-
reaction at the expense, however, of insufficient mixing.
3
.3.5. Effect of disc speed
2
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Attempts were made to improve the mixing and thus, the reaction dependency on residence
time. From Figure 8 and Figure 9, it can be inferred that the product 1 yield and overall
conversion are inversely proportional to the film thickness for flow rates at 3 and 5 mL/s. In the
lower disc speed range, typically between 400 and 1,400 rpm, the film thickness decreases
considerably and the corresponding yield increases significantly. Beyond 1,400 rpm, there is
only a marginal decrease in film thickness from about 30 to 20 µm which is accompanied by a
minimal increase in yield. This seemingly direct correlation between film thickness decrease and
yield increase highlights more rapid mixing and mass transfer taking place across the reduced
path length of thinner films.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
00
100
90
80
70
60
50
40
30
20
10
0
90
80
70
60
50
40
30
20
1ml/s
3ml/s
5ml/s
1ml/s
3ml/s
5ml/s
10
0
0
500
1000
1500
2000
2500
3000
0
500
1000
1500
2000
2500
3000
Rotational Speed (rpm)
Roataional Speed (rpm)
(a)
(b)
Figure 8 (a) Product 1 yield; (b) overall conversion against disc speeds of 400 to 2400 rpm at
different flow rates.
90
80
70
60
50
40
30
20
10
0
1ml/s
3ml/s
ml/s
5
0
500
1000
1500
2000
2500
3000
Rotational Speed (rpm)
16
Figure 9 Estimated film thickness at different disc speed at constant total flow rates.
2
2
ACS Paragon Plus Environment

Page 23 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The convergence of the film thickness profiles at high disc speeds in Figure 9 also matches the
product 1 yield trends seen in Figure 8 (a), suggesting that the limit of highest mixing/mass
transfer has been reached for the process under consideration. Similar convergence profiles were
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
28
observed at high disc speeds in micromixing studies conducted in the SDR. This observation is
1
5
in agreement with the findings made by Hinrichsen that the micromixing segregation index is
directly correlated to the calculated film thickness.
The ortho-lithiation reaction was carried out in three reactors and the operating conditions of
the best runs are summarized in Table 7. The experience gained from the experiments should
provide adequate basis to illustrate the critical differences in performance that would be expected
between batch and flow processes at industrial scale. Table 7 presents the operating conditions
that would be used in the batch and flow process designs.
Table 7 Highest yield obtained in the different reactors and their respective operating conditions.
STR
T-reactor
SDR
17 ˚C
Coolant temperature -80 ˚C
17 ˚C
(˚C)
Yield (%)
96
99
87
Amide
1 0.4
0.4
0.4
concentration (M)
Step
1
combined -
1.2
1.2
flow rate (mL/s)
Residence time (s)
Mins (depends on ~0.4 s
dosing rate)
Micromixing time (s) Seconds
~0.14 s
-4
15
~3 x10 s
0.125-0.02 s
BuLi mole eqv
DMF mole eqv
1.2
1.3
1.2
2
1.2
2
3.4. Design of the stirred tank reactor, T-reactor and SDR reaction systems for 3 tons per year
production.
Having obtained laboratory data for the three process technology options we proceeded to
generate designs for operation at a nominal 3 tons per year scale. For each technology we
2
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
considered whether the laboratory data were representative of what could be achieved at the
larger scale.
For the batch STR, due to lack of suitable large equipment that can provide a proper cooling
system (-80 ˚C), we were unable to perform ortho-lithiation at our designed industrial scale of 28
L. However, it would still be possible to design the STR batch process by making some realistic
assumptions based on the laboratory results at 50 mL Multi-max ™ scale. It was identified that
the reaction is most sensitive to temperature which is directly controlled by the n-BuLi dosing
rate. To ensure the cooling capacity of the reactor is sufficient to cope with the n-BuLi dosing
rate, a conservative overall heat transfer coefficient value is assumed. In addition, good mixing is
required to avoid temperature gradient. The power dissipation of the 50 mL Multi-max ™ at its
maximum achievable agitation speed is used as a basis for providing sufficient mixing required
for the 28 L reactor. The refrigeration method is one of the biggest difference between the
laboratory Multimax ™ and designed industrial process. The vapor-compression refrigeration
system is selected for the large scale cryogenic reactor design which is based on highly idealized
model.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A two-stage selected cascade system schematic is shown in Figure 10. The refrigeration
system consists of two compressor units, two heat exchangers, two throttles and an evaporator is
required to achieve a refrigeration temperature of -80 ˚C. Refrigerants R-32 and R-23 are
selected for the cascade cycle to provide maximum thermodynamic efficiency and recycled after
each batch. An ideal cycle model was assumed, taking into account the compressor in the
cryogenic reactor is driven by an electric motor with an efficiency of 0.72. The utilities
consumed will be mainly for compressor duties, pumping, agitation and chilled water for
2
4
ACS Paragon Plus Environment

Page 25 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
condensation. In order to achieve similar mixing intensity, the agitation power of the large scale
STR is designed to have similar agitation power as the bench scale of 0.053 W/kg.
1
.5 kg Amide 1 per batch
T-01
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Amide 1 13 kg THF per batch
P-01
T-023.6 kg 1.6 M n-BuLi in hexane per batch
THF
Added over 30 min
P-02
T-03
N-BuLi
0.6 kg DMF per batch
Added over 12 min
P-04
P-03
T-05
Product
T-04
DMF
1.65 kg product per batch
4
.4 hours per batch
819 batches per year
3000 kg product per year
3
2 L
1
EP-02
R23 expansion valve
E-01
R23 evaporator
CP-02
R23 compressor
HE-01
R23-R32 heat exchanger
EP-01
R32 expansion valve
2
09 kg chilled water
per batch
Chilled
water
CP-01
R32 compressor
C-01
P-05
R32 condenser
Figure 10 Schematic process diagram for batch process.
For both the T-reactor and the SDR flow processes, their process designs would be based on
the actual laboratory operating conditions (flow rate, temperature, residence time, etc.) used in
their best runs in Table 7 that are able to deliver an annual production of about 3 tons per year.
This provides more realistic operating conditions on the flow reactors, rather than artificially try
to predict the yield at other operating conditions. One point to note is that for all the flow
experiments conducted either in the T-reactor or the SDR, the maximum duration of the run is
about 5 minutes due to volume limitation of the syringe pump. However, the large scale design
requires the run duration to be extended to 8,000 h.
2
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Although there was no clogging in the T-reactor experiments as the runs performed were very
2
9
short, there could be a possibility of agglomeration of the precipitate over time and result in
clogging of the mixing chamber. One possible option to get around this problem is to build in
several T-reactors that can be easily switched to another when pressure build up is sensed. The
clogged T-reactors can then be disposed and replaced swiftly as the T-mixers and tubes are
relatively cheap. In this way, sudden disruption to the process due to clogging and maintenance
of the T-reactor can be avoided entirely.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Even though a small amount of lithium salt was observed on the surface of the disc after the
experiments, the SDR is clearly less prone to clogging. The SDR maintenance can be scheduled
2
9
regularly without the worry of sudden severe pressure drop as seen in the T-reactors. Another
operational advantage of the SDR is the variation in rotational speed which offers an additional
degree of freedom as the disc speed can be controlled independently of the flow rate to achieve
target mixing intensity without affecting the throughput.
One common drawback in the flow processes is the less efficient method to prepare the amide
1
solution as additional equipment are required to dissolve the amide 1 in THF. The respective
reactor setups are shown in Figure 11 and Figure 12. 32 L of amide 1 solution (0.4 M) will be
prepared in a 41 L mixing tank (M-01) in batch mode which will provide 8 hours of amide 1
solution supply for the T-reactor. To ensure continuous supply of the amide 1 solution, two
mixing tanks are required to alternate between each other when one is consumed. Similarly, two
product storage tanks (T-05) will be required.
For the T-reactor system (Figure 11), the utilities used will be mainly the electricity for
silicone oil bath cooling, pumping and agitation for mixing tanks. To avoid evaporation of
butane during the reaction, a back pressure of 2 bars is applied.
2
6
ACS Paragon Plus Environment

Page 27 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
2.7 kg Amide 1 per batch (8 h)
Amide 1
solution
Product 1.4 mL/s
T-01
Amide 1
24 kg THF per batch (8 h)
BPR-01
(
0.4 M) 1
mL/s
Backpressure regulator
P-01
41 L
M-01
P-02
T-05
Product
Mixing Tank
T-02
THF
0
.4 kg product/h
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8
000 hours per year
3040 kg product/year
Circulated silicon oil
17 degC)
(
n-BuLi 0.2 mL/s
P-03
P-04
T-03
N-BuLi
DMF 0.06 mL/s
T-04
DMF
Figure 11 Schematic process diagram for T-reactor flow process.
For the SDR flow process (Figure 12), the utilities used will be mainly the electricity for rotating
the disc, silicone oil bath cooling, pumping and agitation for mixing tanks. No backpressure is
applied in this system, so butane is expected to evaporate from the surface of the disc and which
will be diluted by the continuous supply of nitrogen gas. The mixture of gas then escapes
together with the final reaction mixture through the sink underneath the SDR. Although argon
gas was used in the experiment, nitrogen gas can also be used and is considered in the reactor
design.
2
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2.7 kg Amide 1 per batch (8 h)
T-01
Amide 1
24 kg THF per batch (8 h)
Amide 1
1
mL/s
P-01
41 L
M-01
P-02
Mixing Tank
T-02
THF n-BuLi 0.2 mL/s
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
P-03
P-04
T-03
N-BuLi
DMF 0.06 mL/s
T-04
DMF
Product 1.4 mL/s
TCU
P-05
T-05
Product
.29 kg product/h
0
8000 hours per year (0.85 occupancy)
281 kg product/year
2
Figure 12 Schematic process diagram for SDR flow process
The following general design assumptions were made:
1
) No loss of product 1 yield due to workup, assume all product 1 can be recovered. (workup
is excluded from the study)
2
3
) All reactors are assumed to be air and moisture free during the course of the reaction.
) At the end of the reaction, butane gas (byproduct) is assumed to be diluted and discharged
via nitrogen purge.
4) All chemicals are at ambient temperature which is at 20 ˚C.
5
6
) Concentration of purchased n-BuLi in hexane is always 1.6 M.
) Total number of operation hours per year in a plant is 8000.
7) T-reactor and SDR reach steady state in less than a minute observed from experiment data,
assume insignificant product 1 loss due to unsteady state flow during start up and shut
down.
2
8
ACS Paragon Plus Environment

Page 29 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
8
) No observable degradation of product 1 after addition of DMF during storage based on
GC.
9
1
) Pump efficiency assumed to be 50 %.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0) Specifications of purchased raw materials is assumed as in Table 8.
Table 8 Specifications of the raw materials
Raw materials
Specifications
4-chloro-N,N-diisopropylbenzamide (amide 1) Solid, >99% pure
n-BuLi in Hexane
Tetrahydrofuran (THF)
Liquid solution, 1.6 M
Liquid, anhydrous, 99.9% stabilized with BHT
Dimethylformaide (DMF)
Liquid, anhydrous, 99.8%
3
.5. Comparisons of sustainability performance
Based on the above process designs, we evaluated the sustainability performance in terms of
volume, mass and energy efficiencies and estimated the expected operating (OPEX) and capital
cost (CAPEX) respectively (Table 9). The reactor volume in the flow processes are smaller than
the equivalent batch process because of the shorter residence time. In terms of the footprint of
the major equipment, the T-reactor and SDR process are only 20 % of the batch equivalent.
Maximum processing inventory at any point of time for batch process includes the stirred tank
reactor volume. As for the T-reactor and the SDR flow processes, it refers to the volume of the
reactor and the piping of the reactor system. The mixing tanks are excluded from the process
equipment as they do not deal with n-BuLi and are not involved any reactions; they are instead
considered as part of the storage unit. On this basis, the T-reactor and SDR flow processes
provide much smaller processing inventory as compared to batch process. This would provide
significant safety benefit especially when dealing with hazardous chemical like n-BuLi.
2
9
ACS Paragon Plus Environment